The #1 limitation of an LLM is that it doesn't know what a fact is.
SAFETY
-
Implicit Bias Revealed in GPT-4o and Latest LLMs
By
–
Many benchmarks today fail to observe bias in GPT-4o. But are such models truly unbiased? This new paper reveals implicit bias in the latest LLMs such as GPT-4o and LLama Author @baixuechunzi is on alphaXiv answering questions about her latest work: https://
alphaxiv.org/abs/2402.04105
v2
… -
Comprehensive Review on AI Healthcare Applications Published
By
–
Yet more proof here to add to the comprehensive review by @trishgreenhalgh and colleagues below https://
thelancet.com/journals/ebiom
/article/PIIS2352-3964(24)00192-0/fulltext
… -
Beyond Technocratic Paradigm: Integrating Natural and Social Systems
By
–
“The technocratic paradigm has destroyed that healthy and harmonious relationship…the indispensable need to move beyond that paradigm, so damaging & destructive, will not be found in a denial of the human being, but include the interaction of natural systems w/social systems.
-
OpenAI Creates Safety Committee and Confirms New Model
By
–
[#Article] @OpenAI establishes a Safety and Security Committee and confirms training of a new frontier model https://actuia.com/actualite/open-ai-instaure-un-comite-de-surete-et-de-securite-et-confirme-lentrainement-dun-nouveau-modele-de-frontiere/
… #AI #ArtificialIntelligence -
AI Reliability Paradox: Humans Become More Unpredictable
By
–
the more reliable AI becomes, the more unpredictable I get to be no further explanation will be provided
-
AI Capabilities Advancing Faster Than Evaluation Methods Can Track
By
–
The capabilities are improving so fast that evals can't keep up
-
Periodic AI Model Evaluation Updates and Benchmarking Strategy
By
–
5/ We plan to periodically update these evaluations with: new evaluation sets, to ensure fresh, non-saturated results new models, as they become available Please let us know what you think!
